Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Current Status and Future Promise of the Semantic Web James Hendler Ora Lassila University of Maryland College Park, MD Nokia Research Center Cambridge, MA November 2006 Before We Get Started… • Two views of the Semantic Web: • implementing SEMANTIC applications using web technologies • using semantic technologies to support new WEB applications • Exploring the relationship (and tension) between the two over time • This talk is • a retrospective, a status report • an “interpretation” • some thoughts on the future… 1990’s: “Pre-history” • Rebirth of Artificial Intelligence (end of “AI Winter”) • “big” AI applications • Deep Blue, Mars Rover, Deep Space 1, … • embedded vs. stand-alone • Web AI • IR, statistical NLP, machine learning • lots of data! • Emergence of the Web • new ways of doing things • new business models (even new social models) • new technology • “dot-com” boom • Early forays into “meta-content” Model Complexity Traditional AI applications 1990’s: “Pre-history” dream Exploration of KR applications on the Web ? Applicability Across Domains “The Web” 2000-2001: What Did We Believe? What about the rest of the Web? • Jim: Semantic Web and the advent of pervasive computing (March 2000) DATA AND PROGRAMS 2010 IMAGES AND DOCUMENTS TetherlessDAML The Mobile Semantic Web 2000 1990 SWMU , 03 5 owl.mindswap.org Ge t Info Buy C ar Find Re staurant Q uickTim e™ and a decom pr essor ar e needed t o see t his pict ur e. Buy TV Find Job Buy Airline Ticke t • Jim: Roadmap from the “old” Web to the Semantic Web (October 2001) 2000-2001: What Did We Believe? • Ora: Semantic Web and the advent of pervasive computing (June 1999) • Ora: Roadmap from the “old” Web to the Semantic Web (October 2001) 2000-2001: “Early Years” • “Dot-Com” optimism still prevails: easy to explore new directions • Government meddles with semantics Semantic Web Res. (EU) • DARPA’s DAML program; EU follows • DAML+OIL IST Research efforts Ontoweb Oil Intl Workshops E-Business emphasis • Web community discovers metadata • W3C Metadata Activity • RDF “3-pronged” attack: - DARPA - EU IST - W3C DAML+OIL language EU W3C Members/directors (Dan Brickley , coord) DAML+OIL (webont) Semantic Web Activity W3C www.w3.org/2001/SW DARPA Agent Markup Language RDF-S RDF WebOnt Military emphasis www. daml.org DAML Early Web AI apps Model Complexity 2000-2001: “Early Years” Semantic Web research & new standards (DARPA, EU, …) Joint Committee (DAML+OIL) Early RDF apps (FoaF, RSS) Applicability Across Domains Original Outline (July 2000) 2001 Funded Research WG activity Recommendation • Research, experimentation, early demonstrations • Reminiscent of the early days of the Web Semantic Web Today The Semantic Web of 2002 resembles the early days of the World Wide Web 10 Development funded primarily by Govt, but emerging corporate interest A lot of excitement, but confusion as to business case Open source tools and “geeks in control ” Standards starting to stabilize to point where they permit deployment Developer tools, libraries, languages 10 2003 Funded Research WG activity Recommendation Semantic Web Today “Our” Semantic Web Jan 1, 03: Crawler finds 5.8M+ DAML statements on 20,000+ web pages The Semantic Web of 2002 resembles the early days of the World Wide Web Doesn’t include many instanceKBs tied to ontologies Doesn’t include many v ery large RDFS-based KBs that include some OWL Development funded primarily by Govt, but emerging corporate interest Ontology library at http://www.daml.org has 195 ontologies (March 2003) Open for any one to create A lot of excitement, but confusion as to business case Open for any one to use Open source tools and “geeks in control ” OWL is being supported by large corporation labs Standards starting to stabilize to point where they permit deployment Web tool dev elopers: IBM, HP, Sun, Intel, Fujitsu Content prov iders: Daimler-Chry sler, Nokia, Motorola, EDS,Agfa Developer tools, libraries, languages OWL is starting to be used by thesaurus developers C.f. National Cancer Institute metathesaurus released in OWL Lite United Nations Standard Product Codes av ailable in DAML NASA thesaurus av ailable in DAML Use of semantic markup for Web Services beginning to move beyond basic research 10 DAML-S cited as required reading for Web Serv ices Composition WG 10 Sanken, 03 23 www.mindswap.org 23 • Early government adoption • Emerging corporate interest 2005 Funded Research WG activity Recommendation Semantic Web Today “Our” Semantic Web Jan 1, 03: Crawler finds 5.8M+ DAML statements on 20,000+ web pages The Semantic Web of 2002 resembles the early days of the World Wide Web Doesn’t include many instanceKBs tied to ontologies Doesn’t include many v ery large RDFS-based KBs that include some OWL • C o m p a n ie s g e t t No w in g Development funded primarily by Govt, but emerging corporate interest Ontology library at http://www.daml.org has 195 ontologies (March 2003) Open for any one to create A lot of excitement, but confusion as to business case Open for any one to use Open source tools and “geeks in control ” OWL is being supported by large corporation labs • Commercial tools • I • O • O B Standards starting to stabilize to point where they permit deployment Web tool dev elopers: IBM, HP, Sun, Intel, Fujitsu Content prov iders: Daimler-Chry sler, Nokia, Motorola, EDS,Agfa Developer tools, libraries, languages OWL is starting to be used by thesaurus developers • Lots of open source software • Scalability • H C.f. National Cancer Institute metathesaurus released in OWL Lite United Nations Standard Product Codes av ailable in DAML NASA thesaurus av ailable in DAML • M M r a p e c S • O O N le s u b t n • p o o p u o r r in t a s e s u p c e t s s u p o o ( r K t h e o n t o w R D F S p o r t a L W a b s o p e n im e n t L o p e s o u r a t io n / p 10 Sanken, 03 a c t o R D r i) n - s c e o t u c r c • K o w a r • J e n a , i, a c a d e m R S e é , s a m D F L o lo g F y m in s m e a O c a r la b n a le a g c le t r e m 1 ip 0 le e . 2 s t n o t s r y s t e m e e T r Q I e F u n e g n a A b e c o m ic ib in J o o ls Use of semantic markup for Web Services beginning to move beyond basic research t a P a n y e x p e r S P I in g a v a ila b le f o r u s e , 3 S t o r … e … e DAML-S cited as required reading for Web Serv ices Composition WG • • P r B u ild in g o t c o r é g p o r S a t W e O O d e m P , o n s t O n r a t t o o r ( x x s x ) … b e c o m in g c h e a p a n d 10 e a s y 23 SemTechCo nf , 3 / 05 www.mindswap.org 23 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. k ic ( e T L d im Z d e e W ) ™ t o d a s 2006: You Are Here! Then a Miracle Occurs… Significant Corporate Activity • Semantic (Web) technology companies starting & growing • Joost, Radar Networks, MetaWeb, Siderean, SandPiper, SiberLogic, Ontology Works, Intellidimension, Intellisophic, TopQuadrant, Data Grid, … • Bigger players buying in • Adobe, Cisco, HP, IBM, Microsoft, Nokia, Oracle, Sun, Vodaphone… announcements/use in 2006-2007 • integrator and contractor uptake: Northrop Grumman buys Tucana, Lockheed-Martin uses SiberLogic in FCS, WebMethods buys Cerebra, … • tools being announced: AllegroGraph, TopBraid, … • Government projects in and across agencies • US, EU, Japan, Korea, China, … • Life sciences/pharma an increasingly important market • Health Care and Life Sciences Interest Group at W3C • Many open source tools available • Kowari, RDFLib, Jena, Sesame, Protégé, SWOOP, Onto(xxx), Wilbur, … Significant Corporate Activity 50+ Semantic Web press releases within one month! Growing Government Activity (US&EU) • Agencies moving beyond the "talk" phase • primarily prototyping, but first acquisitions starting • Example: • NASA is developing an enterprise data strategy around using existing data via Semantic Web integration Lots of activities across NASA • Science, Engineering, and Mission all have SWT production or development efforts in place • Now focus in on re-using the data systems we already have in place • Agency wide integration planning is underway for building a federation of models into an integrated information service across all disciplines (A. Schain, 3/06) There's a Lot Out There! Paid ads 2,120,000 hits on "RDF filetype:rdf" 13,600 hits on "ontology filetype:owl" (March, 2006) More OWL Use • The OWL namespace has been declared by 113,000 SWDs (8%) and actually used by 108,000 (7%) • The RDFS namespace enjoys more use, being declared by 677,000 (47%) and used by 538,000 (37%) SWDs owl:Class is the most used term from the OWL namespace with about 1,800,000 instantiations in 68,000 SWDs significant use of two OWL equality assertions: owl:sameAs (280,000 assertions in 17,000 SWDs) and owl:equivalentClass (70,000 assertions in 4,300 SWDs) – their common use may be an indication of increased ontology alignment • • http://swoogle.umbc.edu (from Ebiquity blog, Sept 1, 2006) Semantic WEB Rich metadata Data harvesting & visualization A little Semantics goes a long way Web-based social networks SEMANTIC Web Digital asset management Scientific portals A little Web goes a long way Tools for developers Enterprise Information Integration • Deployment of semantic technologies is easier in a “controlled” environment • such as a corporate intranet • Key benefits from Semantic Web Technology: • reuse of installed clients and servers • careful design of SW languages for Web compatibility • leave data in place, integrate through an RDF store • analogous to 3-tiered Web application • heterogeneity supported by ontologies "Corporate Semantic Web", Gartner "hot pick" for 2006 2006: The Gap Is Closing Model Complexity Semantic Web applications of varying complexity and applicability Applicability Across Domains SEMANTIC Web Lessons • What we learned from AI… • embedded AI succeeded, stand-alone did not • tools are hard to sell • reasoners are a means, not an end • knowledge engineering bottleneck • …applied in the Web context • futureproofing • URIs are important • good standards evolve • languages (RDFS, OWL, RIF, …) • content! Semantic WEB Lessons • Web needed high value sites • personal (homepages, pets) • public (hobbyists, govt) • As these linked up, new functionality emerged • Yahoo, Alta Vista, … • New business models followed… • “give it away” (Netscape) • marketplace (Amazon) • advertising (Yahoo, Google) • Semantic Web? • SHARE; GIVE IT AWAY! • What do we need? • Open Source Tools • Open Source Datasets • Open Source Harvesters The “Layer Cake” is Evolving… (Tim Berners-Lee) 2001 (Tim Berners-Lee) 2006 New Languages Underway • RIF: Rules Interchange Format • representing rules on the Web • linking rule-based systems together • SPARQL: Query language for (distributed) triple stores • the “SQL of the Semantic Web” • GRDDL/RDFa: Integration of HTML and Semantic Web • “embedding” RDF-based annotation on traditional Web pages • OWL: New features, specialized subsets • RDF++/OWL Mini – simplification, identity, scaling to large datasets • OWL 1.1 – additional expressivity for OWL constructs • And more… • multimedia annotation, Web-page metadata annotation, Health Care and Life Sciences (LSID), privacy, etc. Model Complexity Linking Is Power! Applicability Across Domains Semantic Web vs. “Web 2.0” • Data with formal semantics • RDF, OWL • SPARQL, RIF • Spontaneous information integration (finally!) • Semantic Web services, agents • Strong emphasis on open standards • New social phenomena: blogs, wikis, tagging, folksonomies • New user interfaces • AJAX (or: “Rich User Experience”) • “New” kinds of data • microformats, RSS • “mash-ups” • Web services • Plays “fast & loose” with standards Semantic Web & “Web 2.0” • What is their relationship? • Will they stay separate? Does that even make sense? Semantic Web vs. “Web 2.0” Semantic Web & “Web 2.0” • NO! Considerable synergies exist Semantic Web “Web 2.0” Exploiting “Web 2.0” • Vast amounts of “semi-engineered” knowledge • Flickr: tens of millions of keyword-tagged photos • microformatted Web documents • Wikipedia: thousands of carefully documented subjects (in a hierarchy, with disambiguation, …) • Generate “persistent” URIs • ”Tank" http://en.wikipedia.org/wiki/Tank (armament) • ”Tank" http://en.wikipedia.org/wiki/Tank%2C_Pakistan (small town in Pakistan) • Remember: Anything with a URI can be linked to the Semantic Web! Linking of “Web 2.0” & Semantic Web • Using informal Knowledge Engineering (KE) to bootstrap "formal" KE • Extending formal KE from tags/wiki Model Complexity Looking Further Out Applicability Across Domains Where Are the Agents? • “Brave New Applications” • operate autonomously in “unanticipated” situations • exhibit robustness in the face of • changing, inconsistent and unexpected data • variations in reliability, trust • capable of serendipitous behavior, opportunism • Move from the “tool use” of personal computing to systems that work on our behalf • (Semantic) Web services as “plumbing” for agents • emerging as we speak… Pervasive Computing & Semantic Web • Pervasive Computing is an interoperability nightmare! • instead of sometimes connecting a handful of devices, dynamically connect/disconnect/reconnect possibly hundreds of devices • Today, high cost of ensuring interoperation • any interaction has to be specifically designed/engineered • heavy emphasis on application-specific standardization • spontaneous interoperability is next to impossible • The vision is largely contingent on getting unanticipated “encounters” of devices to work • how do you behave in a situation not covered by a standard? • not “future-proof” • Semantic Web is a good match • (it is an “interoperability technology”) Other Emerging Trends • Semantic Web Services • crucial for linking “programs” into the mix • “plumbing” for agents… • Scaling Semantic Web stores to database sizes • Information extraction and semantics ("Web 3.0") • can we “retrofit” semantics on the existing Web? • Semantic Web information creation • can we make it so we don't have to retrofit in the future? • tools that help embed the semantics as a document is created • better dynamic integration of structured data into the Semantic Web • “Semantic Desktop” Summary • Most things we predicted have happened • (or are happening at the moment…) • Some things happened faster than we anticipated • triple store scaling • reasoner performance actually matters • ontologies are there (but very little linking) • Some things are yet to materialize (but we are hopeful) • public information sources (as RDF, OWL, …) • digital convergence, pervasive computing just emerging • little progress on agents Now go out there and make some money off this…! Model Complexity “A Little Web Goes A Long Way” Any Questions? The Semantic Web Applicability Across Domains “A Little Semantics Goes A Long Way”